We have recently installed a CR1000 and AM16/32B into one of our test huts. The AM16/32B is being used to measure 30 Honeywell HIH-4021-003 humidity sensors. The sensors have three legs. The first leg ('+Ve') requires a supply voltage of 5V. The voltage differential is then measured across the remaining two legs ('OUT' and '-Ve'). The literature indicates that there needs to be a minimum load of 80kOhms between the 'OUT' and '-Ve'.
Rightly or wrongly, we ran a lead from the '5V' terminal on the CR1000 into a terminal strip that in turn supplies a constant 5V to all of the sensors. We then used Shortcut to compile a program. We choose the Single Ended Voltage Measurement device and wired the AM16/32B according to the wiring diagram provided by the program. We then ran the program ... but the voltages made no sense at all. In some cases we disconnected sensors and PC400 was still reading the same voltage differentials. After a couple of hours of running different iterations, we left site. Before doing so, we re-wired the AM16/32B such that the 'OUT' lead of each sensor was in the high terminals (e.g. 1H, 2H, 3H, etc.) and the '-Ve' lead was in the ground terminals.
Now that we are offsite, we are trying to figure out if there is any way that we can salvage what we have done with the leads in their current positions before we go back. I guess I'm wondering if there is any way we can use the VOLTSE command to measure the difference between the high terminals and grounds. Is this possible? If so, what would the programming look like?
Thanks for the help ahead of time.
* Last updated by: ZEI2 on 6/21/2011 @ 9:28 PM *
I have no experience with those sensors apart from helping out on a previous post on the forum which did not involve a multiplexer.
When the measurement is being made you need to have a connection made to V+ and the ground terminal on the logger so power can flow to the sensor. I cannot tell from your description but you may have acheived this with your final wiring albeit on every other channel. You would need to check if the ground terminal you have used is connected back to the logger ground as the equipotential ground on the multiplexer is usually but not always connected to ground.
If so you should be able to take the measurements as single-ended voltage measurements ignoring every other reading providing all the sensor power connection are connected all of the time.
One issue that may have caused initial problems would be if you were routing either the 5V or ground connection through the multiplexer relays, such that power was cut to the sensor between measurements. The datasheets for the sensor says it has a 70 ms settling time which means you need to allow at least 70 ms after switching channel before the measurement is made. Shortcut would not have put such a delay in the program for a simple voltage measurement.
One final word of warning these sensors have ratiometric output so if the power supply varies so will the output. The 5V output of the logger is not designed as a precision output (error up to +/-2%), having said that the sensors are relatively poor accuracy anyway so the small errors from the supply may not worry you.
Thanks for the input I really appreciate it.
I took the 5V supply from the CR1000 and am recording it to see that it does not vary too much. For our purposes it is acceptable, but we will keep an eye on it.
One question I have is regarding settling time. You are correct, Shortcut built in a settling time of 3 ms. I will change this to 70 ms as per the cut sheet. However, the way I have wired it up, the sensors see a constant 5V (not a switched 5V). Wouldn't this mean they are always ready to read and as such wouldn't require a settling time?
Finally, do you have a sample of how I would force the CR1000 to read between 1H and ground, 2H and ground, 3H and ground, etc. I am having difficulties understanding how you change what terminals the CR1000 uses to make its measurements. Is this done using the PortSet function? FYI, we have the AM16/32B in the 2x32 mode.
In your final configuration you may have this setup where the sensor is powered all the time so you do not need a long settling time.
In you initial tests you said you used Shortcut which normally uses 4x16 way mode and switches the ground line (allowing a max of 48 sensors). This would need a long settling time. However, you could not force a settling time as long as 70 ms with Shortcut as this needs an extra instruction added to allow this.
TO work out a program suitable for you current configuration please send in the wiring diagram of how the sensors are wired now along with the details above to your local CS office. They should be help you make a simple modification to the program you have - although this is outside the capabilities of Shortcut.